This page has been archived and commenting is disabled.
More On The Mysterious (Ongoing) Quant Disruption
If anyone needs indirect evidence of a "mysterious force" that has been forcing the market away from normal behavior for the seventh month in a row now, one simply needs to look at the dramatic continuing underperformance of quant factors, which in turn force quant models to generate false signals and result in major pain for these traditional allocators of capital and market buffers. Quant models, which rely on a statistically "normal" market, are terminally broken, best visualized by YTD drubbing of traditional market natural and L/S indices. But don't take our word for it. For the latest on this ongoing phenomenon, and for what is probably a good representation by proxy of the mysterious permabid phenomenon (call it however you will), we present excerpts from the latest institutional letter by Matt Rothman, head of U.S. Equity Quantitative Strategies at Barclays.
This month’s performance for quantitative factors was unsettling. All three of our Quantitative Themes markedly underperformed. Market Sentiment was down -2.1%. Quality was down -3.1%. And Valuation was down -3.6%. While we have seen bigger monthly moves in these themes in recent years, the fact that all three of these themes underperformed significantly at the same time is noteworthy. Indeed, this simultaneous underperformance of all our themes suggests to us that a disruption may have been (or still is) occurring in this space.
This sobering thought has me searching for the appropriate words and so I find it necessary to turn to my truest and most reliable muse but I am still struggling to find the most appropriate reference. Is it best to paraphrase the opening of Side 1 of Darkness on The Edge of Town – that is the song, Badlands? Or is better to allude to the eerily appropriate last song of the vastly underrated and misunderstood Nebraska album – that is, Reason to Believe? Or still yet, is it the newest Springsteen song which world-premièred Wednesday night at the Meadowlands, where yours truly was privileged enough to be among the lucky people standing inches from the stage as The Boss belted out, for the first time, Wrecking Ball?
But to be clear, we do not want to belittle the seriousness of the situation. This month marks the 5th out of 7 months that our long/short strategy has underperformed. Our trailing 12 month performance number is sobering even if driven primarily by our performance in April 2009. And that a potential disruption in the space could still be so pernicious with quant aum down as much as we estimate it to be over the past 2+ years, well, that is not comforting either.
And while the current situation is very reminiscent of August 2007 in that it generated major underperformance for quants, at least that period managed to rectify in a very short period of time, granted accompanied by a violent swing in the market. The fact that we are now in month 7 of ongoing quant underperformance, without any notable public implosions should raise red flags. One explanation, the one that Zero Hedge has been promoting, is that as traditional L/S and M/N quants have been redeemed and otherwise slighted, other new players have muscled in and taken their spot: most notably incipient market structure monopolists such as Goldman Sachs.
We have searched for prior occasions where we have seen rolling monthly simultaneous underperformance of a similar magnitude. There were few instances of analogous behavior: August 1958; October/November 1974; May/June 1993; May 2001; October, November and December 2001; November 2002; and May/June 2003. Notably, August 2007 was not a comparable period as the model misbehavior was short-lived and self-corrected quickly. In a number of these periods, there was no active quantitative money management community – computers, while invented, were rare and quite expensive in the early periods – and so it is important to recognize that these disruptions can clearly have a variety of sources.
Or one major one...
And probably the most relevant observation from Rothman, highlighting that even as the status quo presumably persists, the likelihood of major shifts behind the scenes becomes greater and greater. Then again, as BGI is in the process of being transferred to BlackRock, the last thing the financial community needs is awareness of major unwinds occurring under everyone's noses.
Equally importantly, we have still not been able to find any direct evidence of an unwind. Based on our conversations with the Barclays Capital trading desks and with numerous clients, we have no direct evidence to support that an unwind has been happening. Like others, we have seen the press reports that a major pension manager was significantly restructuring their allocations to outside managers, particuliarly portable alpha (quantitative?) managers. We would only be speculating to think this was the cause of quantitative model misbehavior into the quarter end. Yet, this hypothesis does not strike us as entirely unreasonable either. We just don’t know. But we sure can observe our model’s behavior which is certainly abnormal.
As Evidence A of "abnormal behavior" and what are likely major ongoing redemptions, be it in portable alpha or otherwise, please see the YTD performance of the HFRX Equity Market Neutral Index: down -6.3% through 9/30. But at least all other pro cyclical indices are performing well, alas not well enough for them to pass their high water mark yet. Just ask New York headhunters how many hedge funds are actually hiring.
»
- 8385 reads
- Printer-friendly version
- Send to friend
- advertisements -
ah matt, you got played by the squid.
question is: what are gonna do about it?
Tyler,
Not only do I enjoy your posts, the site and the honesty but there was just something else that really made me feel comfortable here....besides the truth.
So few of my trading comrades or sites even know much about Bruce, but you just rang the bell. Have been following him for over 20 years, literally. What a guy. And the E street band cannot be topped.
This is all going to finish in a very ugly way. Just the timing. From the dark side, hopefully someday to the daylight.
The best
Mr. Rothman Sir...
Quants and all others are merely the fleas on the back of the Feds "Dog and Ponzi Show" at this point in time. Everyone is just along for the ride.
BTW the act ends with the dog rolling over.
So the computers doing quantitative trading are good, but the ones doing high frequency trading are bad?
"Then again, as BGI is in the process of being transferred to BlackRock, the last thing the financial community needs is awareness of major unwinds occurring under everyone's noses."
While I am legally not allowed to divulge any details, I can tell you a generalization that I am participating in an undwind of a hedge fund, and the reason it is ongoing (and has been for a year), primarily as a result of remaining assets being illiquid; huge losses will likely be realized on remaining assets.
"Just ask New York headhunters how many hedge funds are actually hiring"
Short RHI.
RHI do lot more than executive recruiting. Besides, as if funadamentals have anything to do with valuation.
It is painfully obvious the machines have taken over. They are exploiting the weaknesses of human nature we are not even aware of. Noone can predict or alter what they do anymore, not even the illuminati. However, we have become so dependent on them we can not turn them off. Realize which part of you is the animal and which part is the computer. Then say goodbye to the first one, silicon people of our glorious planetary mind do not need it. Enjoy the Singularity, Jesus was a transhumanist.
Another reason for market neutral underperformance is higher stock cross-correlations. It's making short cross-gamma hedges more expensive.
I have watched every trade in every major index for the last ten years. Only in the last few months has the movement of prices seem foreign to me. It's not natural but it could be the new normal. This is what happens when you eliminate or weaken all of the competition. We need bsc leh and mer in the game again.
> One explanation, the one that Zero Hedge has been
> promoting, is that as traditional L/S and M/N quants have
> been redeemed and otherwise slighted, other new players
> have muscled in and taken their spot: most notably incipient
> market structure monopolists such as Goldman Sachs.
Tyler, suppose I were to believe that GS is a "market structure monopolist". How pray tell does their activity change market valuations so that L/S and M/N quant strategies break?
Is it your assertion that GS and their brethren running HF market making strategies are actually causing distortions to market pricing? Assuming so since logically that is all I believe you could have meant, what is the mechanism for alterning the market pricing given that their purchases and sales net out to zero - pretty much on a daily basis?
It is quite amazing to me that every move the market makes which does not seem rationale is somehow blamed on GS, JPM, GETCO, and a few other major players...
Perhaps trash is performing better than models would predict because companies with large long date USD debt could see the weight of that anchor alleviated from a USD devaluation....
Perhaps the models are performing poorly because they were constructed poorly....
Perhaps there was an assumed constant in the model which has undergone material change in the last year (which may or may not yet have been identified as a variable)....
Perhaps 3 models over the last 7 months is not all that statistically meaningful....
I just find it silly that everything seems to be tied back to GS, especially when taking a broader view of the world, they are not even remotely as meaninful to valuations as interest rate policies, politics and competitive currency devaluations.
"Circular trading" is evidence of price manipulation. if the machines trade between them to higher levels without concommitent stock delivery, price will move up.. ask any stock manipulator.. Of course you should have the volume, which machines have anyway with the reported 70% of the trading volume by HFT now..
do i get a finance degree if i read all of this and understand 15% :)
I don't believe their purchases and sales always net out to zero. they look at the big picture. its not necessary to profit on every trade. sometimes the price of temporary control,is very low in comparison to benefits generated.this is way bigger than GS, and their strategy involves many aspects, like a chess game. If they do something that doesn't look profitable, you're probably not seeing it right. just my opinion, I'm just a neophyte
Time for Matt Rothman to reflect upon
http://www.youtube.com/watch?v=lrADIyOPC2o
http://www.youtube.com/watch?v=Yogm2hlpnhY
Well pardon me if I don't feel bad for these pricks. These are the same guys who have been rolling mom and pop investors for years. Now they get rolled by a bigger fish and they cry foul?
Front row tickets to Springsteen; ya boohoo. Asshole.
These quant strategies worked well when they were new and kept working well as long as additional money was being allocated to them. The problem is that everyone has the playbook now days. When a bunch of cash was being dumped into the space over the last few years (think of how many pension funds and endowments shifted massive parts of their portfolios into these strategies, for instance), they worked great -- not because they were laws of nature, just because people believed they were, and the new money coming in made it a self-fulfilling prophecy. Now that the allocations to the quant space are flat to down, they no longer have that tailwind and are subject to the slings and arrows of those who would trade against them for profit.
The big blow up in August of 2007 really showed people how vulnerable these strategies were. As the markets rose over the past 2 quarters, a lot of big shops managed to refill their war chests and have been going after these quant strategies -- knowing full well that without new capital coming in the quants are vulnerable. Push them far enough and they will have to throw in the towel, either due to hitting internal risk limits or getting withdrawals from their lps.
He could do better to explain what his L/S strats / calls were and how much have they underperformed in past 7 months.
Then the author and us, could reach conclusions.
He just says that if strats have not worked for 7 months, something is wrong with the markets.
I think that this is a self-involved soliloquy meant to release the pain of somebody who has lost a lot.
Its unreasonable and blind of him to expect that this post is meaningful for anyone, including himself.
My advice to you, the author, is, 'it will work mate, keep up the faith and don't let emotions overpower your rhetoric'
Zerohedge, why are you becoming a garbage chute? Please do not make us read this garbage. Many people are loosing money by being bearish. It does not mean that they have to loose their dignity, reason and presentation.
You aint seen nothin' yet. Just wait until the HFT bots finally un-freeze Megatron Squid. People are too focused on the Markets that they forget the REAL prize - the All-Spark.
My guess is that Optimus Quant will be destroyed by Q1 of 2010 and, once the All-Spark is in the hands of Megatron Squid, the Market will transform into a a million mechanical squid-lets and destroy mankind.
Can a quant program be reverse engineered? Not being a quant, I'm only speculating, but if one knows what signals drive a typical quant program, cannot a program be written to draw them in, then spin them around, then spin them around again. Since so many quants are roughly the same program, couldn't one player lead them all in circles? Maybe this is what Sergey late of GS had his hands on?
All edges run their course. That's what trading is. Stop crying about your quant strategies not working and start thinking of new strategies to exploit the weakness' of those who are exploiting yours.
Marge N Call,
Makes sense to me; I like the way you think.
"Market Sentiment was down -2.1%. Quality was down -3.1%. And Valuation was down -3.6%. "
WTF are "sentiment" "quality" and "valuation"?
I thought the point of Black Scholes Merton was to remove the made - up indicators, not to loop them back in at a higher semantic level?
Just me, or this guy is talking gobbledygook? (and if so, who gives a flying fuck about what model he's using)
There is no conspiracy here.
The problem with Rothman's factors are that they are primarily backward looking. The market has aggressively discounted an economic recovery before it has become apparent. This plays havoc with models such as these as stock prices have risen on potential and not what they have necessarily achieved.
He needs the market to fall and disappointment with the economic recovery to set in for his models to make back some of their losses. Alternatively, better thought out Factors and Factor allocation would have helped.